Contractile analysis with kriging based on MR myocardial velocity imaging

Abstract

Diagnosis and treatment of coronary artery disease requires a full understanding of the intrinsic contractile mechanics of the heart. MR myocardial velocity imaging is a promising technique for revealing intramural cardiac motion but its ability to depict 3D strain tensor distribution is constrained by anisotropic voxel coverage of velocity imaging due to limited imaging slices and the achievable SNR in patient studies. This paper introduces a novel Kriging estimator for simultaneously improving the tracking and dense inter-slice estimation of the myocardial velocity data. A harmonic embedding technique is employed to determine point correspondence between left ventricle models between subjects, allowing for a statistical shape model to be reconstructed. The use of different semivariograms is investigated for optimal deformation reconstruction. Results from in vivo data demonstrate a marked improvement in tracking myocardial deformation, thus enhancing the potential clinical value of MR myocardial velocity imaging. © 2008 Springer-Verlag Berlin Heidelberg.